@kevalprajapati· yesterday
One-shot debugging on project_litmus
project_litmus
DEEP WORK
25.8x your typical session
4.1M tokens vs typical 160.5K
4.1Mtokens
74prompts
46:37:12time
3,674lines
Claude Codeclaude-opus-4-8Python$101.05
mostly building
19h 24m session running commands and testing. 64K tokens across 1 prompts using claude-sonnet-4-6.
Tokens / prompt56.0k
Cost / line$0.028
Cache hit97%
Burn rate$2.17/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 66%Verifying 26%Thinking 7%
Comments